Premium
TIME SERIES ANALYSIS OF FURROW INFILTRATION
Author(s) -
Nasseri Abolfazl,
Neyshabori Mohammad Reza,
Fard Ahmad Fakheri
Publication year - 2013
Publication title -
irrigation and drainage
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.421
H-Index - 38
eISSN - 1531-0361
pISSN - 1531-0353
DOI - 10.1002/ird.1756
Subject(s) - autoregressive integrated moving average , autoregressive model , mathematics , infiltration (hvac) , autocorrelation , statistics , series (stratigraphy) , logarithm , time series , stationary process , mathematical analysis , physics , meteorology , geology , paleontology
Measured cumulative infiltration ( Z ) over time usually shows significant correlations with their contiguous values. This autocorrelation in a series restricts application of regression analysis to determine the coefficients of empirical equations. In such cases, time series analysis is preferred to process the modeling. In this study, the possibility of furrow infiltration modeling by time series techniques was investigated based on experimental data obtained from infiltration measuring at 56 sites over a field by blocked furrow infiltrometer. Results revealed that the Z data were nonstationary in the mean and variance. Logarithmic transformation and the first‐order differencing were applied to convert the measured series to the stationary ones. Cumulative infiltration data were successfully analyzed as an autoregressive integrated moving average (ARIMA ( p , 1, q )) process. The autoregressive terms had a more momentous role than moving averages in modeling Z series. Results confirmed that the obtained models explain more than 99% of the variability in the calibrated values. The ARIMA ( p , 1, q ) parameters were modeled as a function of flow or soil characteristics. The first and second parameters of autoregressive terms (φ 1 and φ 2 ) and moving average terms (θ 1 and θ 2 ) were modeled as a function of independent variables. Comparative results showed that performance of time series modeling was better than commonly used infiltration equations such as that of Kostiakov to characterize furrow infiltration. Copyright © 2013 John Wiley & Sons, Ltd.